Four Words for Friend.

Steven Poole has a brief Guardian review of what sounds like it might be an interesting book, Four Words for Friend by Marek Kohn:

The author, a native Polish speaker, makes a powerful case for knowing more than one language as a life-enriching skill that may enlarge our sympathies in a world that wants to build walls. Though, as Kohn unsentimentally points out, linguistic differences can sometimes be erected as walls themselves. In Papua New Guinea, home to 800 languages, one village decided to change its word for “No” so as to be different from its neighbours.

We learn much here about the politics of languages in Latvia, India and the US, as well as the science of language acquisition in infancy and adulthood, and the pros and cons of growing up perfectly bilingual. Surprisingly, it was the expert consensus only half a century ago that this was harmful to intellectual development, but current research suggests the opposite.

Alas, it seems to be infected with Sapir-Whorf Syndrome (“different languages, because they carve up the world in different ways, cause speakers to perceive and think differently […]. Hence the book’s title: in Russian, one is obliged to specify one of four levels of closeness when referring to a friend”), but that’s a venial sin that the reader can correct for. Thanks, Lars!


  1. I’m Russian and I’m struggling to identify the four levels of closeness this book asserts. There’s друг (“friend”), близкий друг (“close friend”), and знакомый (“acquaintance”). Sure there’s also товарищ (“comrade, colleague”) and приятель (“buddy”) but I’m not sure either is very common these days. Where I grew up (the Moscow topolect, give or take) we only used the друг/знакомый distinction. Overall this system doesn’t seem very different from, say, English. From the pragmatics point of view, of course there are cultural specifics: e.g. Americans tend to use “friend” for relationships that Russians would normally place in the “acquaintance” category, but I used to chalk it up to the same grade inflation trend that brought us “awesome” instead of “okay,” “amazing” in place of “typical,” and all the self-enhancements familiar to anyone who’s ever screened resumes at international companies. But I digress. My point is that I suspect this system of levels of closeness in Russian, even if there are indeed four of them, may not be so special.

  2. I agree with all your points; I would have griped about the four levels, but I figured if I were going to do that I should take the trouble to find out what Russian words or phrases he had in mind, and I was too lazy.

  3. David Eddyshaw says

    John McWhorter (I feel like I’m invoking a powerful demon merely by mentioning his name, for some reason) has sensible remarks somewhere on this “weakened Sapir-Whorf” bit, boiling down to “yes, language does influence how you conceptualise the world, but to such a minor degree that for practical purposes it might as well not.”

  4. David Eddyshaw says

    Apropos of absolutely nothing whatsoever, and purely because it is a joy to give praise where it is due, I note the following Correction in this week’s Economist:

    Johnson’s previous column mistakenly said that “Tuesday” includes the Indo-European root dyeu twice. Dyeu produced the word for day in other languages, but English “day” is thought to be from dhegwh, “to burn”. Sorry.

  5. David Marjanović says

    “yes, language does influence how you conceptualise the world, but to such a minor degree that for practical purposes it might as well not.”

    It has a really dramatic influence on color perception, though.

  6. David Eddyshaw says

    McWhorter in fact discusses this very example in extenso; his conclusion is indeed that the differences are real but trivial. (It’s in The Language Hoax.)

    The reliability of the linked article is clear from e.g.

    As languages evolve, their ability to describe the world gains complexity … the first description of color to emerge in a young language is the differentiation between light and dark.

    The author has no idea what he is talking about.

  7. David Marjanović says

    Of course “young language” is nonsense, but isn’t it fair as a description of the Berlin-Kay theory?

    Anyway, that’s not what I meant. Keep going till you reach the picture with all the squares arranged in a circle. In Standard Average European, they’re all green, and all almost the same shade of green… while in some other languages they’re two different colors as different as green and blue to us. Knowing that, how long does it take you to see it?

  8. J.W. Brewer says

    I must say that it doesn’t bode well if the illustrative anecdote used for the book’s title doesn’t survive native-speaker factchecking.

  9. Judging from the book’s cover art Russian friends are of one of the blue, red, yellow, or green variety. Inside the book, the “four words for friend” factoid is sort of attributed to Jacobson (“On linguistic aspects of translation”, which discusses some Russian words, but nothing about friends). I found this blog post, which discusses 4 words for friend in Russian, but Gregory already explained all that there is to explain.

  10. ktschwarz says

    The pictures with the green and blue squares arranged in a circle are fake. There was no such experiment. Debi Roberson did do the study comparing how English and Himba children learned color words, as described in the first four paragraphs after “Language Impacts How We See”. She also did a study with squares arranged in a circle and picking the one that’s a different color, but it was with Korean speakers, not Himba; it involved not cyan and green, but yeondu (yellow-green) and chorok (green); and the pictures on the Day Translations site are not from Roberson’s paper, they are “dramatizations” cooked up for the BBC Horizon show. See Language Log for details. We’re in Radio Yerevan territory here. Unfortunately, the BBC version has been spread all over the place.

  11. David Marjanović says


    Well, thanks, I’ve updated my bookmarks and will have to read the actual papers at some point.

  12. Well, I’m glad I posted it so some much-needed debunking could be done!

  13. I actually sent the link because I thought the treatment of bilingualism sounded sympathetic. I hope the guy isn’t as weak on the concepts there, because the message deserves more exposure.

  14. Kristian says

    Why is the Sapir Whorf hypothesis so attractive to people? Even if it were true that Russians divide their friends/acquaintances into four categories, I would say that was a part of Russian culture, not something their language makes them do.

  15. Gregory forgot “собутильник”.

    On the other hand I can’t think of an adequate Russian word to translate “pal” – which to me has become a jocular or even purposely insincere word for friend. (e.g. “Your Plastic Pal Who’s Fun To Be With” – Douglas Adams).

  16. The only thing that comes to mind is дружок, but that’s probably not very close.

  17. David Eddyshaw says

    the treatment of bilingualism sounded sympathetic

    Quite so. His heart is undoubtedly in the right place. I wish there were a good way of stopping people adducing bad arguments for good causes. If anything, it undermines the good cause by making it wrongly appear as if depends entirely on bad arguments.

    I have quite a few ideological allies that I secretly wish might see their way to defecting to the opposition …

  18. J.W. Brewer says

    David E.: The next step beyond wishing they might defect is dark speculation that they have indeed already done so and are now secretly on the payroll of the Other Side, tasked with discrediting your cause.

  19. How about кореш / корефан? Doesn’t rescue the 4-types nonsense, but gives you one more flavor. The real friend flavor which is missing but would be useful is “an acquaintance who makes you believe in some nonsense factoid”

  20. Heh. Wiktionary gives “crony, pal, buddy, dawg (close or old friend).”

  21. David Eddyshaw says


    Fortunately my ideological foes are not that sophisticated …

    Wait … perhaps all this time they’ve just been pretending to be unsophisticated? The cunning devils! That’s just the sort of thing they would do.

  22. Wiktionary gives “crony, pal, buddy, dawg (close or old friend).”

    Etymology is rather questionable, but apparently a 1923 slang dictionary is consistent with wiktionary’s explanation.
    Still it is just as likely that it came from Hebrew, like a number of other classic slang words of the epoch.

  23. Israeli Hebrew uses two words for ‘friend’: חָבֵר/חָבֵרָה xaver/xavera and יָדִיד/יְדִידָה yadid/yedida (m/f). There’s also מַכָּר/מַכָּרָה makar/makara ‘acquaintance’, which corresponds closely to the English.

    I can’t describe well the division between xaver and yadid. The former is closer, to be sure. Since xaver/a also means ‘boyfriend/girlfriend’, yadid/yedida is used when this ambiguity is present and undesirable. Yadid is a hair more formal, and children seldom use to describe their friends. Even though you can say either xaver tov or yadid tov ‘a good friend’ to emphasize the closeness, the latter still implies a little more separation. In any case, I can’t think of a case where you couldn’t refer to a xaver as a yadid or vice versa. It’s a pragmatic, not a semantic difference.

    Xaver is also used for ‘comrade’ (in socialist talk), or member of some organizations (including political parties and governmental bodies).

    A recent regrettable development is the word יָזִיז yaziz, which is what in American English is called ‘a friend with benefits’. That comes from replacing the daleth in yadid with the visually similar zayin, the mere letter whose unexpected presence suggests sex at its slangiest. I first encountered the word when reading it in a not completely obvious context. I went from non-comprehension to aha in about ten seconds. Israeli slang revels in this kind of sly reference.

  24. English cater cousin means ‘close friend’ or ‘distant relative’. Its etymology is contentious. I have only discovered the word just now.

  25. David Marjanović says


    Also suspected to be the origin of Viennese /ˈhavɐrɐ/ “buddy/pal”.

  26. John Cowan says

    Huh. It just seems obvious to me that cater cousin is similar to cater corner > kitty corner, someone in the opposite corner of a genealogical chart from you.

  27. What I don’t understand is the semantic shift from ‘distant relative’ to ‘close friend’.

  28. Why would the Yiddish /x/ become Viennese /h/?

  29. David Marjanović says

    the semantic shift

    Extensions from “brother” to “friend” via “fellow member of something” aren’t uncommon.

    Why would the Yiddish /x/ become Viennese /h/?

    Once upon a time, [h] was the word-initial allophone of (short) /x/.* To this day, word-initial /x/ is absent from many, and word-initial [x] or [χ] absent from all, German dialects except Pretty High Alemannic, and from all Standard accents; loans with such are borrowed with /h/ or /k/.

    Why /h/ and not /k/ in this particular instance, I don’t know. But /h/ seems a phonetically better choice to me at least.

    Edited to add: we gave a /k/ to Chruschtschow. Part of the reason must be that the other chr- words are all Greek loans and all get a /k/ as well; part of the reason for that is probably the fact that initial [x] isn’t native.

    * I eventually stopped myself from going into the whole history of German back fricatives… tell me if you’re interested.

  30. The Annotated Hobbit has a discussion of what exactly Gandalf means by the word when he tells Beorn that Radagast is his “cousin,” concluding that a literal familial relation is probably not intended.

  31. ‘Brother’ > ‘friend’, sure. ‘Distant relative’ ditto is odd.

    Are there other dialects close to Yiddish (which is not Alemannic) which have initial [x]?

  32. Ellen K. says

    Don’t be distracted by the words “distant” and “close” in the English translations of the term. Think about the meaning. People one step removed from your immediate family. Doesn’t seem at all odd to me for a language to have a term that means both.

  33. David Marjanović says

    Are there other dialects close to Yiddish (which is not Alemannic) which have initial [x]?


    In Yiddish itself, /x/ doesn’t occur in words of Germanic origin either.

  34. David Marjanović says

    …initial /x/, of course.

  35. Owlmirror says

    A recent regrettable development is the word יָזִיז yaziz, which is what in American English is called ‘a friend with benefits’. That comes from replacing the daleth in yadid with the visually similar zayin, the mere letter whose unexpected presence suggests sex at its slangiest.

    Isn’t the fact that there’s some similarity to forms of the verb zuz; “move”, relevant?

    I see that Wiktionary has the term, but does not support my suggested etymology.

  36. I wonder if he gives dues to Anna Wierzbicka, who wrote about the four words for ‘friend’ and compared them with three other languages in Understanding Cultures Through Their Key Words: English, Russian, Polish, German, Japanese. I love the methods she brings to that discussion, looking, for instance, at the proverbs and idioms about friendship and how they fit with the different ‘friend’ conceptualizations.

  37. Having read the book (I reviewed it dilettantishly here), I think Kohn does get a little Neowhorfian but he also understands the objections to it and his interpretation of the phenomenon is sometimes subtler than it seems – one of the points he makes is that if polyglots feel that they become more voluble in Spanish, calmer in German (or whatever) – which many of them do – then, in fact, they really do tend to act and speak that way, so in a sense it’s a self-fulfilling prophecy. Obviously this is something that only really applies to people who speak more than one language, not to native speakers within a monoliguistic language community, but then that’s the whole point of the book.

  38. Why would the Yiddish /x/ become Viennese /h/?


    Borrowed from Yiddish חבֿר‎ (khaver), from Hebrew חבר‎ (khaver, “friend”).

    IPA(key): [ˈhɒvɛr]
    Hyphenation: ha‧ver
    Rhymes: -ɛr
    haver (plural haverok)

    1. (slang) pal, buddy, dude

  39. David Marjanović says

    Going through Hungarian would immediately explain the /h/, but Hungarian words are not that common in Viennese, and the [ɒ] would have been borrowed either as such, i.e. au, or as the other vowel in that region (which corresponds to Standard German /a(ː)/, but not as /a/ (Standard German ä or ei, mostly).

Speak Your Mind